Towards realistic 3D elastic models of Canadian channel and reef structures

Gary F. Margrave, Samantha Taylor, Joanna Kathleen Cooper

We discuss the creation of two detailed elastic models of Canadian stratigraphic exploration targets and present initial views of wavefield simulations through these targets. Our first model, which is a buried channel sequence beneath a stratified overburden, is more mature than our second model, which is a pinnacle reef beneath a stratified over burden. The process used to create both models was similar. We began by creating geologically plausible maps at five depth levels within the channel system or the reef. The maps were intended to be similar, but not identical, to known structures. Seven distinct lithologies were identified for each model, and these maps specified polygonal regions, which were each assigned one of the lithologies. The maps were then digitized and, for each map, three duplicates were created specifying Vp, Vs, and density as uniform properties within 2D polygons. Then a specially developed 3D kriging algorithm was used to interpolate a number of new maps between each digitized one. This process also created plausible property gradients within the polygons on the original maps. The resulting channel structure contains 13 levels, each described by a 241x241 grid, while the reef structure contains 24 levels on a somewhat larger grid. An overburden and underlying succession, which vary only vertically, were then crafted for each model from smoothed representations of available well logs. The final 3D models specify Vp, Vs, and density at all points in the 3D volume. The channel sequence is 120 m thick at a depth of 1 km and with horizontal apertures of 2.7 km in each direction. The reef model is slightly larger and deeper. Sample 3D shots have been simulated at several positions over the channel model using our Rayleigh-Sommerfeld algorithm and using a commercial finite-difference code. The simulations show that our models produce a very complex elastic response that can be used to evaluate the performance of imaging algorithms and to study footprint effects.